For mapping regions in the United States, JMP comes with map files for states and counties. In a previous blog post, Mike Vorburger showed how to add a map files for US Metropolitan Statistical Areas. Sometimes I hear requests for ZIP code map files, and now they are available for download from the JMP File Exchange (free SAS login required to access files).
Why did it take so long? ZIP codes are messy for several reasons:
- ZIP codes are designed for mail delivery and don’t always correspond to two-dimensional geographic areas. Some are government agencies or buildings or streets.
- Some geographic areas do not have ZIP codes, such as military bases and national parks.
- ZIP codes change frequently, which can cause data and shapes to be out of sync.
- ZIP code shape files are huge, making them relatively expensive to draw.
- Until recently, shape files from the US Census Bureau came in many disjointed files.
While the first three issues aren't going away, performance is becoming less and less of a problem, and the disjointed download is no longer an issue.
The US Census Bureau now provides a single shape file for all ZIP codes. The download is 800 MB, but I’ve reduced the detail level, and the resultant JMP map files are (only) 60 MB. For comparison, our US county map files are under 1.5 MB, so that’s still a lot of data. I should point out that these aren’t really ZIP codes – they’re ZIP Code Tabulation Areas, which is the Census Bureau’s approximation of ZIP codes.
The above map shows ZIP codes colored by land area rank, so bigger ZIP codes are blue and smaller ones are red. No surprise that smaller ZIP codes occur in cities. (This post is about the map, not the data!) The white areas are those not covered by ZIP codes. I had to use Right-click > Graph > Customize to hide the shape outlines so they didn’t dominate the small areas. I also added the background US state boundaries, which is noticeable for Alaska.
Zooming in to Florida you can see the smaller ZIP codes clustered around the coast. Setting the background states to fill with green, you can readily see the Everglades National Park and other areas not covered by ZIP codes.
Zooming further into the Tampa Bay area, you can see how small the ZIP code regions can get in a city like St. Petersburg.
In practice, ZIP code maps are more useful for local areas, where you can distinguish each shape. The main benefit of using ZIP codes for mapping is the small granularity which can make it easier to see localized patterns.
I've set the bar low for insightful data. What kind of data is good for showing by ZIP code?